maximum concave convex tolerance in lathe bed ways

[michiganbuck]

I think those numbers are darn chose for a lathe
(reading every 8": (converted from mm so the numbers are a little strange)
(0 / 0 / +.00025 / +.00037 / +.0005 / +.00037 / +.00037 / +.0005 / +.00025 / 0 )
I would let it rest for a long time, and then if still out ..00025 I would make a center foot and jack it up .00025 (or what).

 

[Tyrone Shoelaces]

Another fact that shows that this was not the case is that the lower part was grinded together and when you measure it with a micrometer it is equal along the entire length... if I had removed more material from the middle it would have been thinner.
the measurement obtained (in my humble perception) shows a constant arch (there is no part where the slope regresses, just the constant rise)

I’m not so sure about that. If all the faces were ground at the same time shouldn’t they be parallel to each other anyway. How much did you have taken off ?

On another issue I always insisted that there was a “ witness “ left on the ways by the grinder. Just a small unground area. That told me the minimum amount had been takeoff by the guy.

Regards Tyrone

 

[gbent]

You are showing a lathe that needs leveling. Your numbers should sum to zero. Then lets talk about the shape of the way.

 

[Richard King]

The bed plate is steel and it bolts to the bottom bed casting. Hardinge sold new plates or you could send the bed plate back to the factory where the would regrind them if they weren't worn to deep. As they are flame hardened. They ground them on way grinders and ground the bottom and side dovetails first operation and then flipped it over and ground the top. They used coolant. I know this as I toured the plant in Elmira and saw them building new machines. The wood floors looked like a bowling alley.

 

[Hbjj]

You are showing a lathe that needs leveling. Your numbers should sum to zero. Then lets talk about the shape of the way.

Absolutely agree 100%

Sounds like he has wood?!!! Under the leveling feet....

Ps IMHO the sag he is seeing is the expected result of grinding on a long relatively thin section like this

 

[Miyamura]

I’m not so sure about that. If all the faces were ground at the same time shouldn’t they be parallel to each other anyway. How much did you have taken off ?

On another issue I always insisted that there was a “ witness “ left on the ways by the grinder. Just a small unground area. That told me the minimum amount had been takeoff by the guy.

Regards Tyrone

I don't know if I understood correctly (my English is terrible)
but guides were grinded together both at the top and bottom, so that when measuring with a micrometer the thickness is the same from end to end on both the flat part of the seadle and the TS

 

[Miyamura]

think those numbers are darn chose for a lathe
(reading every 8": (convert

I think those numbers are darn chose for a lathe
(reading every 8": (converted from mm so the numbers are a little strange)
(0 / 0 / +.00025 / +.00037 / +.0005 / +.00037 / +.00037 / +.0005 / +.00025 / 0 )
I would let it rest for a long time, and then if still out ..00025 I would make a center foot and jack it up .00025 (or what).

The result of these numbers are forming an "almost perfect" curve when I put them on a graph... I've never seen anything like that.
I measured it again today and the numbers didn't change anymore (I even used another level of precision as a stubborn measure)
I lifted the middle of the bed with the crane, very lightly, to see what happens

 

[Miyamura]

Absolutely agree 100%

Sounds like he has wood?!!! Under the leveling feet....

Ps IMHO the sag he is seeing is the expected result of grinding on a long relatively thin section like this

they are just beams screwed to the concrete so that I can better place a steel plate with the threaded hole and have support for the lifting screws... I use this instead of vibrastops (I don't know what you call it, but it's the part of the Photograph )

 

[Miyamura]

Anyway, today the person who did the grinding came here.
He helped me find the center of the problem and wait for it to come back, but he has already committed to redoing the job if it doesn't work.
One of his theories (in addition to the twisting and tension that has accumulated over the years) is that when the bed was placed on the grinding machine due to inattention there may have been a lot of tension and after releasing the bed from the table it returned to normal but now with a curve.
but at this point it's all speculation
In the years I've worked with this, I've never caught a case like this.

 

[michiganbuck]

QT Op: ((reading every 8": converted from mm so the numbers are a little strange)
(0 / 0 / +.00025 / +.00037 / +.0005 / +.00037 / +.00037 / +.0005 / +.00025 / 0 )

With these numbers on a lathe I would not try to grind out the error.
One would be taking a big chance to have - numbers.
A spread of .00025 is darn close for a lathe, I would be in favor of the whole bed being + 00025.

Just the way a lathe bed sets on a grinder and then sets on a shop floor can be a variance of a few tenths.

 

[Hbjj]

Just curious but what is the manufacturer recommended installing method

It's almost always in the service manual.

It's usually J shaped threaded rod set in X inches of concrete providing a solid foundation that would allow to easily adjust this sag out .

This is pretty much industry standard on this type of tool

I know some modern machines can use the vibra stop type pad but they usually have a much deeper section

 

[Miyamura]

Just curious but what is the manufacturer recommended installing method

It's almost always in the service manual.

It's usually J shaped threaded rod set in X inches of concrete providing a solid foundation that would allow to easily adjust this sag out .

This is pretty much industry standard on this type of tool

I know some modern machines can use the vibra stop type pad but they usually have a much deeper section

The machine is very old (1957) but I have the manual.
recommends using vibrastop or attaching it to the floor (both cases with concrete of at least 4"). the way I did it is as if it were attached directly to the concrete. to give it more stability I attach a very thick wooden beam to the floor with several screws and on the wood I place a layer of steel for the regulators to rest on and also to thread the screws for the lathe bases
I'm more and more certain every day that it was my fault for leaving it sitting for so many years with terrible storage :/
It's the only machine I have that has this problem.
I even have two metal planes fixed in the same way and they are great.

 

[Hbjj]

It sounds like you've answered your own question then.

Good luck!

 

[michiganbuck]

Dry grinding is possible but to grind a big part like a lathe bed to an acceptable surface finish would be near impossible with not having part heat swelling, wheel wear or part distortin from the set -up..Even wet to <.0003 would be tough.
I think grinding a lathe bed to a spread on.00025 would be near impossible or difficult wet or dry.
When down to .0003 hand finishing is much safer.
And the effect on the part is not the same as only bed height variance when you get to <.0003 in height on the flat way..

 

[Miyamura]

Dry grinding is possible but to grind a big part like a lathe bed to an acceptable surface finish would be near impossible with not having part heat swelling, wheel wear or part distortin from the set -up..Even wet to <.0003 would be tough.
I think grinding a lathe bed to a spread on.00025 would be near impossible or difficult wet or dry.
When down to .0003 hand finishing is much safer.
And the effect on the part is not the same as only bed height variance when you get to <.0003 in height on the flat way..

michiganbuck, due to the difference in language I don't know if I understood exactly what you said (I rely a lot on Google Translate)...
but what I understand is that at this length a variation smaller than 0.0003 is impossible?
so an acceptable difference at 78"?
please correct me if I misunderstood
In fact, the topic strayed a little from the question I had, which is precisely this... I didn't find any literature on the greatest variation in the total length of the lathe bed acceptable.
In the books and references I have, they only talk about the difference every 12" (in this aspect my lathe is excellent)

 

[michiganbuck]

Simply saying that the variance you find in that one aspect may not cause a part variance equal to the same when turning a part on the machine. Turning a part will be equal to the error on the flat way and on the V way like an average.
The way a large part like a lathe bed sets on a grinder chuck .and then sets bolted to a base or set on a shop floor can make some accuracy changes because of the way it sets a little differently.
Grinding or scraping an 80" bed to zero error, and then adding the weight of the saddle and tail assembly likely may make some error.
A variation of .0003 could be taken out, corrected by scraping or honing easier than grinding.
Wet grinding such a large part would be best because just one wheel pass on a long part can cause some heat swelling and a possible burn.
If the lathe is low at center a center foot might be added like a jack screw.

 

[Richard King]

Lathes new from the factory that are flame hardened or soft iron. Specifications are .0002" in 12" or 1 foot. You machine new machine tolerance a little over .0015"

You see higher tolerances on super precious machine that are inside a temperature controlled room (air conditioning and heater kept at 68 degrees F. Your lathe was not designed to hold tenths. It would be foolish to try to get tenths. Buck is correct!

 

[Miyamura]

Lathes new from the factory that are flame hardened or soft iron. Specifications are .0002" in 12" or 1 foot. You machine new machine tolerance a little over .0015"

You see higher tolerances on super precious machine that are inside a temperature controlled room (air conditioning and heater kept at 68 degrees F. Your lathe was not designed to hold tenths. It would be foolish to try to get tenths. Buck is correct!

exactly!
I'm not looking for this level of precision, because I don't need it, I just want to leave the best possible within the limits stipulated by a rule... as I said, I didn't find any reference on the total flatness tolerable for this type of machine in the total length .
It's very easy to go crazy with this kind of obsession, so a real reference helps put a limit on the madness lol... but I want to do it as well as possible.
The reality is that I worked with my leblond for 10 years with wear on the ways above 0.05" and was able to make parts with very low tolerances.
Yesterday I did something that I should have done from the beginning and out of stupidity I didn't remember to do it... I contacted the factory (it still exists )
I'm waiting for the answer (support couldn't tell me either )

 

[Miyamura]

I managed to speak to the factory engineer!
tolerance in current manufacturing is 0.0004" by 39.5"
(0.01mm per linear meter)
at the time of manufacture mine was .0008" by 39.5" (0.02mm per linear meter
this is the tolerance to leave the factory

 

[michiganbuck]

I have an air blast cooler that brings the part temperature down to near freezing. I would not trust it for grinding a long grinding pass part *but perhaps the op’s grinder person has perfected a method to grind dry. “The method that works, works”.
I have envisioned gridding methods and machine designs that could likely accomplish this.
A tool bit/insert rise and fall of .002” going + and - of part center. oo1" would not change the part diameter that same amount. My rough guess the variance on a 4” diameter part might be .000003 or so. (A pm member might calculate the for us as i may be off by a zero place ?.)